
Couple of weeks back, Andrea Di Maio of Gartner wrote a blog post arguing that cloud computing is killing open source in the government. He says that the government agencies are finding cloud based alternatives to proprietary infrastructure, OS, office applications, etc. instead of the traditional open source software. According to him, this implies death for open source in the government. The arguments he offers in support of his thesis are as follows
- One of the main drivers of open source traction was vendor independence and since there are not many departments in the government who want to download open source applications and use it without the paid support from the vendor, this vendor independence factor becomes meaningless. So, according to him it is no more an attractive factor for the customers.
- The other main driver for open source, cost, is also irrelevant because the government agencies has to pay for support to the vendors and, now a days, it is much cheaper to procure from the cloud providers than pay these open source vendors for support.
I see many gaps in his arguments and I thought I will use this blog post to highlight these gaps and then link to my previous take on this topic.
First and foremost is the way he defines vendor independence. The vendor independence factor in any open source discussion was never meant to imply taking out the vendor completely from the equation. It is more about the choice of using the software without relying on the vendor for anything (installation, maintenance, customization, modification, (even) upgrades, etc.). It is about using the software even after the demise of the vendor (a fact he has highlighted but not emphasized in his post). I don’t think that the argument against open source, using the reliance on vendors, is a strong one. In fact, a government or an enterprise need not rely on the vendor for support to the open source products. There are third party vendors who offer support for open source software at a much lower cost. The initial days of Oracle Unbreakable Linux did just that. They took Redhat Enterprise Linux and offered their support package on top of it. If any customer wants cost savings on support packages, they can take the help from the ecosystem of reliable third party vendors supporting various open source products/projects. The perception that you need to include the vendor responsible for the development of open source software to get support is not true. There are few advantages to using the software developers for support (depending on their business model) but it need not be the case.
I agree with him partially on his cost argument. I do agree that the economics of scale in the case of cloud computing is too tempting to embrace. I do agree that the cost savings offered by cloud computing to the government agencies are enormous and it makes complete business sense to use cloud computing in government. I have emphasized this advantage many times in this space including the recent post about the Apps.Gov announcement. However, I don’t agree with the implication that the cost factor in cloud computing puts open source at a disadvantaged position. I think it is an apples to oranges comparison. The cloud economics doesn’t single out open source software alone. Rather, it uproots the entire traditional software marketplace, including both the open source and proprietary software. It is a complete paradigm shift in the way we consume computing and its impact is not just felt by the traditional open source software but, also, the proprietary ones.
Having said that, I will link back to my take on why and how open source is still relevant in a cloud based world. There are two ways in which we can approach the question of relevancy of open source in the cloud based world (Okay, three but I am not including Richard Stallman’s approach in the discussion here). One school of thought, advocated by people like Tim O’ Reilly, is asking us to ignore the licensing aspect and focus only on open architecture, open protocols, open formats, etc.. The other school of thought, to which I subscribe, favors the idea of keeping open source in focus while, also, emphasizing on aspects like open architecture, formats, interoperability, data portability, etc.. There is a strong reason to emphasize open source within the cloud realm and I have given my reasons in this post, along with an argument put forward by Simon Wardley of Canonical.
Now, let us also consider the fact that not all governments will be comfortable putting their data on the public clouds at this point of time. It could lead to the proliferation of the so called Government Clouds (G-Cloud in the case of UK) and, in some cases, even private clouds within a government department. The realistic possibilities of such a federated cloud ecosystem within government itself is a good reason why open source software isn’t going anywhere. Even if we consider the case of
SaaS applications replacing traditional open source desktop applications, we have to take into account the fact that the open
source software is at the foundation of many of these SaaS vendors and without the OSS, these vendors cannot offer the cost savings that differentiate them from the traditional software vendors. Then, there is always the opportunity of open source being the endgame for the SaaS vendors going out of business.
To conclude, open source isn’t going anywhere, whether it is in the government or the enterprises. If we see open source as a philosophical platform instead of a business model (which it isn’t) or a developmental model (an oversimplification), it is quite easy to see that open source will continue to play its role irrespective of how we consume the computing resources.